Super-Resolution Microscopy with High Spatial Precision

Top

00:00:11.12 00:00:12.01 Hello everyone!
00:00:12.02 00:00:16.09 Welcome to the microscopy lecture series.
00:00:16.10 00:00:17.23 My name is Xiaowei Zhuang.
00:00:17.24 00:00:20.15 And I am a professor at Harvard University
00:00:20.16 00:00:22.18 and an investigator of
00:00:22.19 00:00:25.05 Howard Hughes Medical Institute.
00:00:25.06 00:00:27.17 And today what I am going to tell you about
00:00:27.18 00:00:30.00 is a brief introduction of
00:00:30.01 00:00:32.21 super-resolution fluorescence microscopy,
00:00:32.22 00:00:36.02 which breaks the diffraction limit.
00:00:36.03 00:00:39.12 And such imaging methods allow us
00:00:39.13 00:00:42.19 to look at cells with much better clarity.
00:00:42.20 00:00:44.21 For example, here we see a conventional
00:00:44.22 00:00:46.18 image of mitochondria.
00:00:46.19 00:00:49.01 And here is a three dimensional
00:00:49.02 00:00:51.27 super-resolution image of mitochondria.
00:00:51.28 00:00:53.15 So I'll tell you a few methods
00:00:53.16 00:00:55.07 that allow us to get there.
00:00:55.08 00:00:59.13 But first let me tell you why we like
00:00:59.14 00:01:02.00 fluorescence microscopy so much.
00:01:02.01 00:01:05.02 As we all know, this is one of the
00:01:05.03 00:01:07.24 most widely used imaging modality
00:01:07.25 00:01:10.21 for biological research.
00:01:10.22 00:01:12.12 And two of the advantages
00:01:12.13 00:01:14.10 of fluorescence microscopy
00:01:14.11 00:01:18.15 is really very well recapitulated by this movie that
00:01:18.16 00:01:20.10 I downloaded from the book
00:01:20.11 00:01:22.11 .
00:01:22.12 00:01:23.29 by Bruce Alberts et al.
00:01:24.00 00:01:28.04 So first right away jumping into the scene
00:01:28.05 00:01:30.27 even without even knowing what it is
00:01:30.28 00:01:33.16 is things that are moving, changing,
00:01:33.17 00:01:35.21 highly dynamic all the time.
00:01:35.22 00:01:37.19 And what you actually see here are
00:01:37.20 00:01:40.00 actually diving embryonic cells.
00:01:40.01 00:01:42.25 The blue is the staining for DNA
00:01:42.26 00:01:44.25 and the green is microtubule.
00:01:44.26 00:01:49.05 So because light is really quite non-invasive,
00:01:49.06 00:01:54.29 it actually allows us to monitor living samples,
00:01:55.00 00:01:58.17 even not only living cells, but also living organisms,
00:01:58.18 00:02:01.03 with relatively little perturbation.
00:02:01.04 00:02:03.07 And because of that,
00:02:03.08 00:02:05.04 and because of light microscopy
00:02:05.05 00:02:07.13 have reasonably fast time resolution,
00:02:07.14 00:02:10.20 so we can actually catch dynamics
00:02:10.21 00:02:13.29 as is happening in cell in real time,
00:02:14.00 00:02:16.04 as one important advantage.
00:02:16.05 00:02:17.22 The other advantage is aslo
00:02:17.23 00:02:19.15 quite obvious from this movie,
00:02:19.16 00:02:22.02 and that is you see different color objects.
00:02:22.03 00:02:25.25 Because light is really wonderfully colorful,
00:02:25.26 00:02:28.05 and also there are so many different
00:02:28.06 00:02:30.03 colorful fluorescent probes these days,
00:02:30.04 00:02:35.16 such as fluorescent proteins, fluorescent dyes,
00:02:35.17 00:02:37.04 quantum dots, you name it.
00:02:37.05 00:02:41.06 And also because they are
00:02:41.07 00:02:44.19 chemically or biochemically very specific approach
00:02:44.20 00:02:47.26 that allow us to specifically attach these probes
00:02:47.27 00:02:50.03 to the molecule of interest,
00:02:50.04 00:02:53.24 so when you combine these three aspects together,
00:02:53.25 00:02:56.18 what we have is a tool that allows us to look at
00:02:56.19 00:02:59.07 different gene products inside the cell
00:02:59.08 00:03:02.20 simultaneously with high molecular specificity
00:03:02.21 00:03:05.19 that when combined with live cell imaging
00:03:05.20 00:03:09.20 allow us to study molecular processes in real time.
00:03:09.21 00:03:11.24 And hence we can learn a lot about
00:03:11.25 00:03:13.10 what's going on inside the cell
00:03:13.11 00:03:16.08 from fluorescence microscopy.
00:03:16.09 00:03:18.27 That said, light microscopy also has
00:03:18.28 00:03:20.19 a quite significant downside.
00:03:20.20 00:03:22.19 And that is its resolution.
00:03:22.20 00:03:25.12 Traditionally resolution of light microscopy
00:03:25.13 00:03:28.02 is limited to a few hundred nanometers.
00:03:28.03 00:03:30.07 While that might sound small,
00:03:30.08 00:03:33.15 things inside the cell are much smaller.
00:03:33.16 00:03:35.25 For example, protein molecules are often
00:03:35.26 00:03:37.25 just several nanometers in size.
00:03:37.26 00:03:40.06 And not only that,
00:03:40.07 00:03:41.22 cellular environment is
00:03:41.23 00:03:43.20 a very crowded environment,
00:03:43.21 00:03:46.15 with molecules packed in high density.
00:03:46.16 00:03:48.23 So ideally if we want to see the
00:03:48.24 00:03:50.27 interactions between molecules
00:03:50.28 00:03:53.03 we would need a resolution
00:03:53.04 00:03:55.14 that is up to the molecular scale.
00:03:55.15 00:03:57.28 So before I tell you
00:03:57.29 00:04:01.28 how we try to approach that kind of scale,
00:04:01.29 00:04:05.03 in other words, to recapitulate this introduction,
00:04:05.04 00:04:07.06 we would like to keep the advantages
00:04:07.07 00:04:09.03 of light microscopy,
00:04:09.04 00:04:10.22 the live cell imaging capability,
00:04:10.23 00:04:13.27 the multicolor molecular specific imaging capability,
00:04:13.28 00:04:17.19 but push the resolution towards the molecular scale.
00:04:17.20 00:04:20.05 What's preventing us previously is
00:04:20.06 00:04:23.07 this so-called diffraction limited resolution.
00:04:23.08 00:04:25.13 Now since Jeff Lichtman is going to
00:04:25.14 00:04:28.12 give another iMicroscopy lecture
00:04:28.13 00:04:30.18 that will tell us what exactly
00:04:30.19 00:04:32.24 is diffraction limited resolution,
00:04:32.25 00:04:34.26 so let me just go through that very briefly.
00:04:34.27 00:04:36.15 Because light is a wave,
00:04:36.16 00:04:38.22 all waves are subjective to diffraction.
00:04:38.23 00:04:42.26 That means if we try to use a lens to focus light
00:04:42.27 00:04:45.22 no matter how an objective we can find
00:04:45.23 00:04:49.28 the focal spot nonetheless will have a finite size.
00:04:49.29 00:04:51.26 Even with the best high numerical
00:04:51.27 00:04:54.03 aperture objective in the world,
00:04:54.04 00:04:57.26 the focal spot size in the lateral dimension
00:04:57.27 00:04:59.19 is about 200 nm,
00:04:59.20 00:05:02.06 and along the axial direction
00:05:02.07 00:05:03.12 where light propagates down
00:05:03.13 00:05:05.13 is even approaching 1 um.
00:05:05.14 00:05:09.03 Now if we use such a spot to scan our sample,
00:05:09.04 00:05:12.03 and then each of these small features
00:05:12.04 00:05:14.13 will be broaden by the spot.
00:05:14.14 00:05:16.02 That's why the spot is also called
00:05:16.03 00:05:17.28 the point spread function.
00:05:17.29 00:05:19.14 If the features,
00:05:19.15 00:05:22.22 let's say two of these features are widely separated
00:05:22.23 00:05:25.16 despite each one of them is broadened,
00:05:25.17 00:05:29.11 none the less we can resolve them.
00:05:29.12 00:05:31.12 And if the features get closer,
00:05:31.13 00:05:34.00 closer, and closer together
00:05:34.01 00:05:38.22 eventually the two images of these two features
00:05:38.23 00:05:41.13 will overlap so substantially
00:05:41.14 00:05:43.06 so that you just see a single blob,
00:05:43.07 00:05:44.17 and that's when you no longer
00:05:44.18 00:05:47.03 are able to resolve them.
00:05:47.04 00:05:49.17 That means the resolution is basically
00:05:49.18 00:05:51.04 comparable to the spot size,
00:05:51.05 00:05:55.23 which is about 200 nm in the x, y direction
00:05:55.24 00:05:58.27 and about 500 nm in the z direction.
00:05:58.28 00:06:00.11 And this was first recognized
00:06:00.12 00:06:02.21 in the 1870s by Ernst Abbe,
00:06:02.22 00:06:05.24 hence diffraction limit is also called the
00:06:05.25 00:06:07.13 Abbe resolution limit.
00:06:07.14 00:06:12.02 Now with this principle of resolution limit in mind,
00:06:12.03 00:06:15.14 there are currently a couple of
00:06:15.15 00:06:18.15 different classes of methods
00:06:18.16 00:06:22.21 that allow us to break the diffraction limit.
00:06:22.22 00:06:28.10 The first is to take advantage of pattern illumination
00:06:28.11 00:06:31.09 to effectively shrink the size
00:06:31.10 00:06:33.01 of the point spread function.
00:06:33.02 00:06:38.19 And this class include methods like
00:06:38.20 00:06:41.06 Stimulated emission depletion microscopy,
00:06:41.07 00:06:43.29 or known as STED microscopy,
00:06:44.00 00:06:48.09 or its generalized form of RESOLFT microscopy.
00:06:48.10 00:06:50.10 And as well as the
00:06:50.11 00:06:52.23 structured illumination microscopy,
00:06:52.24 00:06:54.29 and particularly the saturated form of the
00:06:55.00 00:06:58.01 structured illumination microscopy,
00:06:58.02 00:07:01.14 as S(SIM) also allow us to get resolution
00:07:01.15 00:07:04.20 that is substantially beyond the diffraction limit.
00:07:04.21 00:07:07.16 The second class of method
00:07:07.17 00:07:11.08 take advantage of single molecule imaging
00:07:11.09 00:07:14.20 and stochastic switching of single molecules,
00:07:14.21 00:07:19.17 and representative examples include
00:07:19.18 00:07:22.25 the Stochastic optical reconstruction microscopy,
00:07:22.26 00:07:24.11 the short name is STORM,
00:07:24.12 00:07:28.00 Photo-activated Localization Microscopy [PALM],
00:07:28.01 00:07:30.22 and Fluorescence Photo-activation
00:07:30.23 00:07:32.25 Localization Microscopy,
00:07:32.26 00:07:35.13 the in short name of, that is, FPALM.
00:07:35.14 00:07:39.01 So I'm going to in this lecture briefly
00:07:39.02 00:07:42.14 elaborate both classes of approaches.
00:07:42.15 00:07:45.06 There are also these other
00:07:45.07 00:07:47.25 single molecule based methods
00:07:47.26 00:07:52.13 that take advantage of other ways to
00:07:52.14 00:07:56.09 change the single molecule signaling time
00:07:56.10 00:07:59.10 for example, through molecular bindings,
00:07:59.11 00:08:02.00 through fluctuation of fluorescence signal,
00:08:02.01 00:08:04.19 and through bleaching and so on.
00:08:04.20 00:08:09.20 These approaches include methods like
00:08:09.21 00:08:14.27 PAINT, SOFI, BALM and generalized SHRIMP,
00:08:14.28 00:08:19.06 and 3B analysis, as well as other methods.
00:08:19.07 00:08:23.28 But due to the limited time that we have in
00:08:23.29 00:08:26.16 this short introduction lecture,
00:08:26.17 00:08:29.24 I don't have time to elaborate on these methods,
00:08:29.25 00:08:33.00 but would refer interested audience
00:08:33.01 00:08:34.26 to these references
00:08:34.27 00:08:38.00 and other references so that you could find.
00:08:38.01 00:08:40.25 So first let me give you a brief
00:08:40.26 00:08:43.12 introduction of STED and RESOLFT.
00:08:43.13 00:08:45.07 Back in the 90s,
00:08:45.08 00:08:50.08 Stefan has already the foresight to realize that
00:08:50.09 00:08:51.25 it is actually possible to
00:08:51.26 00:08:54.12 break the diffraction limit in the far field.
00:08:54.13 00:08:57.03 And in his elegant STED design,
00:08:57.04 00:09:00.23 he uses two beams to accomplish that goal.
00:09:00.24 00:09:03.01 One of the beam excites fluorescence
00:09:03.02 00:09:04.28 to give this excitation spot,
00:09:04.29 00:09:07.21 and the other is a doughnut-shaped beam
00:09:07.22 00:09:14.01 that induces stimulated emission.
00:09:14.02 00:09:15.23 And that deplete fluorescence
00:09:15.24 00:09:18.20 at the edge of this excitation spot,
00:09:18.21 00:09:20.27 and effectively shrink the point spread function
00:09:20.28 00:09:22.21 to a much smaller size.
00:09:22.22 00:09:24.22 So one could imaging that if you use such
00:09:24.23 00:09:27.25 a small sized spot to scan your sample,
00:09:27.26 00:09:29.29 you'll get a resolution that is much
00:09:30.00 00:09:31.15 beyond the diffraction limit.
00:09:31.16 00:09:36.22 Later on, he, Stefan,
00:09:36.23 00:09:37.29 also generalize this approach
00:09:38.00 00:09:40.15 to the RESOLFT form,
00:09:40.16 00:09:43.13 which actually take advantage of
00:09:43.14 00:09:47.04 a general fluorescence transition,
00:09:47.05 00:09:48.21 reversible fluorescence transition,
00:09:48.22 00:09:52.14 that just limited to stimulated emission
00:09:52.15 00:09:54.05 to shrink the point spread function
00:09:54.06 00:09:58.09 and then give us the sub-diffraction limit image.
00:09:58.10 00:10:02.09 And this can be extend to the 3D form
00:10:02.10 00:10:05.04 by using a STED beam that also
00:10:05.05 00:10:09.07 deplete fluorescence at the edge of the
00:10:09.08 00:10:11.28 excitation spot in the z direction,
00:10:11.29 00:10:16.06 give you almost isotropically small sized
00:10:16.07 00:10:19.00 fluorescence spot that is much much
00:10:19.01 00:10:20.22 smaller than the diffraction limit.
00:10:20.23 00:10:25.03 And then here are some beautiful 3D STED images
00:10:25.04 00:10:28.00 of mitochondria, and you can compare the
00:10:28.01 00:10:32.04 conventional confocal fluorescence image
00:10:32.05 00:10:35.13 of mitochondria, and the 3D STED images,
00:10:35.14 00:10:38.14 and see a much improved resolution.
00:10:38.15 00:10:42.05 And STED can not only be applied to
00:10:42.06 00:10:46.16 living cells, but is also applicable to living animals.
00:10:46.17 00:10:51.17 Here is a remarkable movie of, STED movie of
00:10:51.18 00:10:55.29 neurons in the brain of a living mouse.
00:10:56.00 00:11:00.12 Since Stefan Hell himself will also give a lecture in
00:11:00.13 00:11:03.22 this iMicroscopy series, so
00:11:03.23 00:11:07.19 I won't further elaborate this approach.
00:11:07.20 00:11:12.13 Now let's go to this other approach, which also use
00:11:12.14 00:11:14.17 pattern illumination to effectively
00:11:14.18 00:11:16.23 shrink the point spread function.
00:11:16.24 00:11:20.05 And this is the structured illumination microscopy
00:11:20.06 00:11:23.02 developed by Mats Gustafsson,
00:11:23.03 00:11:26.21 Dave Agard, Rainer Heintzman and co-workers.
00:11:26.22 00:11:28.27 Now in this approach,
00:11:28.28 00:11:35.25 a sinusoidal excitation pattern is applied to the sample.
00:11:35.26 00:11:38.04 This excitation pattern
00:11:38.05 00:11:41.19 beats with the sample information
00:11:41.20 00:11:46.06 and brings those high spatial frequency information
00:11:46.07 00:11:48.04 that is otherwise not resolvable
00:11:48.05 00:11:50.19 by a conventional microscopy
00:11:50.20 00:11:54.22 into a range that is resolvable by light microscope.
00:11:54.23 00:11:57.09 So for example, what you see here
00:11:57.10 00:12:02.17 is the sample pattern that is superimposed with these
00:12:02.18 00:12:06.03 vertical stripes of excitation pattern.
00:12:06.04 00:12:08.29 And that generate these small fringes
00:12:09.00 00:12:11.17 that has a much lower spatial frequency
00:12:11.18 00:12:13.29 that can be detected by the microscope.
00:12:14.00 00:12:17.25 And with this, we can also have a
00:12:17.26 00:12:20.09 better resolution and diffraction limit.
00:12:20.10 00:12:22.03 Now in the linear form,
00:12:22.04 00:12:25.19 it can improve the resolution by two fold,
00:12:25.20 00:12:29.04 and in the saturated form or nonlinear form,
00:12:29.05 00:12:31.07 we can actually get the resolution much
00:12:31.08 00:12:33.00 beyond the diffraction limit.
00:12:33.01 00:12:34.06 For example,
00:12:34.07 00:12:38.29 what you see here is a 50 nm resolution
00:12:39.00 00:12:41.02 image of fluorescence beads,
00:12:41.03 00:12:44.15 compared to conventional images
00:12:44.16 00:12:47.27 again you see a much higher resolution.
00:12:47.28 00:12:53.05 SIM and also in principle, saturated SIM,
00:12:53.06 00:12:55.10 can also be extended
00:12:55.11 00:12:57.11 to the third dimension to allow us
00:12:57.12 00:13:00.25 to take a 3D supper-resolution image.
00:13:00.26 00:13:05.13 And here what you see is a 3D linear SIM image of
00:13:05.14 00:13:07.24 nuclear pore complexes,
00:13:07.25 00:13:11.23 and in comparison to the conventional image
00:13:11.24 00:13:13.17 you can see these pore complexes
00:13:13.18 00:13:15.04 much better resolved.
00:13:15.05 00:13:18.09 Dave Agard will also give
00:13:18.10 00:13:21.07 a lecture in this iMicroscopy series
00:13:21.08 00:13:24.00 to talk about SIM, so I also won't
00:13:24.01 00:13:27.26 elaborate on this method further.
00:13:27.27 00:13:32.07 So another class of supper-resolution
00:13:32.08 00:13:34.28 fluorescence imaging method
00:13:34.29 00:13:38.11 takes advantage of single molecule imaging.
00:13:38.12 00:13:41.21 So as we know in the single molecule field that
00:13:41.22 00:13:45.09 even though the image of a single molecule is still
00:13:45.10 00:13:46.13 diffraction limited,
00:13:46.14 00:13:49.00 meaning that the width of the image
00:13:49.01 00:13:51.18 is still a couple hundred nanometers wide,
00:13:51.19 00:13:53.29 nonetheless we can localize
00:13:54.00 00:13:56.10 the center of the image with
00:13:56.11 00:13:58.26 a much higher precision,
00:13:58.27 00:14:00.21 and this precision is roughly equal to
00:14:00.22 00:14:01.21 the width of the image
00:14:01.22 00:14:04.27 divided by the numbers of photons
00:14:04.28 00:14:08.09 that we direct take a square root.
00:14:08.10 00:14:10.29 So for example, if we detect 10,000 photons
00:14:11.00 00:14:14.05 from this images, square root of 10,000
00:14:14.06 00:14:17.24 would be a 100, so if we do 200 divided by 100
00:14:17.25 00:14:20.03 that gives us 2 nm.
00:14:20.04 00:14:21.12 So in other words,
00:14:21.13 00:14:24.23 despite the diffraction limit image
00:14:24.24 00:14:28.29 we can localize the centroid position of this image
00:14:29.00 00:14:30.15 with a few nanometer
00:14:30.16 00:14:33.05 or even higher localization precision.
00:14:33.06 00:14:35.27 This is a very powerful approach
00:14:35.28 00:14:37.23 that has already been used
00:14:37.24 00:14:41.10 to study a variety of biological systems.
00:14:41.11 00:14:42.26 For example
00:14:42.27 00:14:45.21 Paul Salvin and co-workers have applied to
00:14:45.22 00:14:50.23 directly visualize hand-over-hand
00:14:50.24 00:14:54.10 movement of kinesin molecules.
00:14:54.11 00:14:57.06 However, powerful as it is,
00:14:57.07 00:15:01.02 this high precision localization
00:15:01.03 00:15:03.27 of single molecules does not directly
00:15:03.28 00:15:07.19 translate into sub-diffraction limit resolution.
00:15:07.20 00:15:10.10 and that is because it requires well
00:15:10.11 00:15:13.16 isolated images of individual molecules.
00:15:13.17 00:15:15.05 One could argue that
00:15:15.06 00:15:18.01 if you already have two molecules
00:15:18.02 00:15:21.12 the images are well isolated from each other
00:15:21.13 00:15:23.13 then they are already resolved by
00:15:23.14 00:15:25.06 conventional imaging approaches.
00:15:25.07 00:15:28.13 So the question of resolution then comes in
00:15:28.14 00:15:32.09 as how do we then separate overlapping images
00:15:32.10 00:15:34.08 of single molecules.
00:15:34.09 00:15:36.29 In particular, you know
00:15:37.00 00:15:41.23 many biological samples are very heavily stained
00:15:41.24 00:15:44.12 with fluorescent probes so that
00:15:44.13 00:15:47.08 their individual images of
00:15:47.09 00:15:48.22 single molecules are heavily
00:15:48.23 00:15:50.06 overlapped with each other.
00:15:50.07 00:15:53.05 So in this scenario, how do we separate the
00:15:53.06 00:15:55.17 images of single molecules and
00:15:55.18 00:15:57.15 localize their centroid position.
00:15:57.16 00:15:59.05 Several years ago,
00:15:59.06 00:16:01.16 we and others had come up with the idea
00:16:01.17 00:16:06.27 of taking advantage of photoswiching of fluorophores
00:16:06.28 00:16:08.22 to break the diffraction limit
00:16:08.23 00:16:11.04 and separate these single molecules images.
00:16:11.05 00:16:15.26 Many fluorescent probes or fluorophores
00:16:15.27 00:16:17.02 are photoswitchable,
00:16:17.03 00:16:20.23 means that they have an ineffective dark state
00:16:20.24 00:16:22.04 and a fluorescent state
00:16:22.05 00:16:24.17 and they can be switched from dark state
00:16:24.18 00:16:26.17 to the fluorescent state by light.
00:16:26.18 00:16:28.01 In that case,
00:16:28.02 00:16:31.28 if we use a weak enough activation light
00:16:31.29 00:16:33.04 and what we can accomplish
00:16:33.05 00:16:37.01 is then activate only a small subset of the molecules
00:16:37.02 00:16:38.08 in the field of view,
00:16:38.09 00:16:41.22 such a small subset so that they are
00:16:41.23 00:16:43.03 isolated from each other,
00:16:43.04 00:16:46.09 and that will allow us to pinpoint the center position
00:16:46.10 00:16:49.12 of these images and determine the molecular positions
00:16:49.13 00:16:53.19 for those activated fluorophores in this particular frame.
00:16:53.20 00:16:56.09 And then we can iterate this process
00:16:56.10 00:16:59.28 over time, and then accumulate
00:16:59.29 00:17:02.09 a sufficient number of
00:17:02.10 00:17:04.26 localizations of individual molecules
00:17:04.27 00:17:09.14 to build up this superresolution image.
00:17:09.15 00:17:11.09 So you can see that in this case,
00:17:11.10 00:17:13.15 the image resolution is
00:17:13.16 00:17:17.08 no longer limited by the diffraction,
00:17:17.09 00:17:20.10 but instead limited by how precise
00:17:20.11 00:17:22.28 we can localize each individual molecule
00:17:22.29 00:17:25.17 and the localization density.
00:17:25.18 00:17:29.13 Now because we use a wide field illumination
00:17:29.14 00:17:34.00 to activate the individual molecules
00:17:34.01 00:17:35.06 in the field of view,
00:17:35.07 00:17:38.06 and there is no patterned illumination
00:17:38.07 00:17:41.03 so the activation process is stochastic,
00:17:41.04 00:17:42.14 you know we can say we activate
00:17:42.15 00:17:45.18 0.1 % of the molecules, but we don't know which
00:17:45.19 00:17:47.29 0.1%, so we termed this method
00:17:48.00 00:17:52.05 stochastic optical reconstruction microscopy,
00:17:52.06 00:17:54.22 and hence this acronym STORM.
00:17:54.23 00:17:57.06 I should also mention that
00:17:57.07 00:18:01.14 in the mean time, two other teams have
00:18:01.15 00:18:05.22 independently developed similar approach
00:18:05.23 00:18:09.11 and published their work almost simultaneously,
00:18:09.12 00:18:14.16 and that's, one of them is the team lead by Eric Betzig
00:18:14.17 00:18:17.06 and Jeniffer Lippincott-Schwartz,
00:18:17.07 00:18:19.20 and they termed their approach PALM,
00:18:19.21 00:18:24.02 and the other team is lead by Samuel Hess,
00:18:24.03 00:18:27.01 and he called his approach FPALM.
00:18:27.02 00:18:30.16 I should also note that
00:18:30.17 00:18:33.13 many other labs have contributed
00:18:33.14 00:18:39.13 to the development of different photoswitchable probes,
00:18:39.14 00:18:42.02 and different photoswitchable mechanisms
00:18:42.03 00:18:45.21 that have made this approach very versatile
00:18:45.22 00:18:48.26 that can be applied to many biological systems.
00:18:48.27 00:18:51.17 However again, due to the limit of time,
00:18:51.18 00:18:55.18 I won't have time to elaborate these different
00:18:55.19 00:19:00.01 switching mechanisms and different switchable probes.
00:19:00.02 00:19:02.20 I'll just give you one example of
00:19:02.21 00:19:04.13 the application of STORM
00:19:04.14 00:19:07.00 to image microtubules.
00:19:07.01 00:19:09.05 As you can see that in this case, this is
00:19:09.06 00:19:11.22 a wide field image of a microtubules
00:19:11.23 00:19:13.20 so where all the fluorophores are turned on,
00:19:13.21 00:19:18.21 so it give you a smeared poorly resolved image.
00:19:18.22 00:19:22.26 Now if we activate a small subset of molecules at a time,
00:19:22.27 00:19:25.13 and then because they are well isolated,
00:19:25.14 00:19:26.27 we can pinpoint their positions,
00:19:26.28 00:19:30.01 we plot their positions at the right hand side.
00:19:30.02 00:19:32.13 And then after enough accumulation,
00:19:32.14 00:19:36.16 we can see this type of STORM images of microtubules.
00:19:36.17 00:19:39.07 And you can see in comparison with the conventional
00:19:39.08 00:19:40.26 image of the same field of view,
00:19:40.27 00:19:45.13 there is a drastic boost of resolution.
00:19:45.14 00:19:48.21 And we also expand this approach
00:19:48.22 00:19:51.25 into the third dimension by taking advantage
00:19:51.26 00:19:55.25 of the sheet of individual molecule images
00:19:55.26 00:19:57.20 to determine their z position.
00:19:57.21 00:20:00.14 Now one of the simplest approach to do that
00:20:00.15 00:20:03.17 is to use the Astigmatism imaging.
00:20:03.18 00:20:08.03 In this case, a cylindrical lens is inserted into
00:20:08.04 00:20:11.08 the detection path, and because cylindrical lens
00:20:11.09 00:20:13.11 only bend light in one direction,
00:20:13.12 00:20:15.03 and not the orthogonal direction
00:20:15.04 00:20:18.05 and therefore individual molecules images are
00:20:18.06 00:20:21.04 in general electrical, and from the electricity
00:20:21.05 00:20:23.01 we can determine the z position of
00:20:23.02 00:20:25.14 individual molecules.
00:20:25.15 00:20:27.11 And then again from the centroid position we
00:20:27.12 00:20:29.18 can determine their x, y position.
00:20:29.19 00:20:32.20 So hence we can determine x, y and z, all 3D
00:20:32.21 00:20:36.14 dimensional coordinates simultaneously,
00:20:36.15 00:20:40.04 and using this approach we first demonstrated
00:20:40.05 00:20:42.27 3D superresolution imaging.
00:20:42.28 00:20:47.20 And this is a 3D STORM image of microtubules
00:20:47.21 00:20:49.22 taken that way.
00:20:49.23 00:20:52.20 And since then there are actually
00:20:52.21 00:20:56.24 quite a number of three dimensional localization
00:20:56.25 00:21:00.15 approaches that have been used for
00:21:00.16 00:21:03.03 3D superresolution imaging.
00:21:03.04 00:21:05.22 For example, the Biplane approach
00:21:05.23 00:21:09.23 the engineering of point spread function,
00:21:09.24 00:21:12.14 in particular, the double=helical point spread function,
00:21:12.15 00:21:16.15 using interferometry to determine z position,
00:21:16.16 00:21:19.26 and the using of micromirror to sort of flip
00:21:19.27 00:21:23.18 the z dimensional information into the x y plane
00:21:23.19 00:21:24.06 and other methods.
00:21:24.07 00:21:28.05 Again, because the limit of time, I won't have
00:21:28.06 00:21:30.21 time to elaborate these methods,
00:21:30.22 00:21:36.05 but to refer interested audience to these references
00:21:36.06 00:21:37.27 and other interesting papers
00:21:37.28 00:21:39.28 that have come out recently.
00:21:39.29 00:21:43.22 So here is an example of
00:21:43.23 00:21:46.15 using a Astigmatism imaging
00:21:46.16 00:21:52.12 to do 3D STORM imaging of actin in the cell
00:21:52.13 00:21:55.27 and actin is, because it's very small dimension,
00:21:55.28 00:21:57.17 it's high packing density,
00:21:57.18 00:22:02.09 one of the actually more difficult objects to be resolved
00:22:02.10 00:22:05.15 by light microscopy, and here you can see that
00:22:05.16 00:22:08.02 even though in the conventional image we don't see any
00:22:08.03 00:22:10.04 individual actin filament resolved,
00:22:10.05 00:22:13.15 but these filaments are resolved very well
00:22:13.16 00:22:17.10 in those 3D superresolution, 3D STORM image.
00:22:17.11 00:22:21.13 And the localization precision we accomplished
00:22:21.14 00:22:27.06 in this particular work is about 4 nm in x,y position,
00:22:27.07 00:22:30.09 and 8 nm in z position.
00:22:30.10 00:22:34.12 And that translate into a resolution measure,
00:22:34.13 00:22:37.14 meaning how far the two molecules need to be
00:22:37.15 00:22:40.01 in order for them to be unambiguously resolved,
00:22:40.02 00:22:45.03 in a resolution to be about 10 nm in x y,
00:22:45.04 00:22:47.23 and 20 nm in z.
00:22:47.24 00:22:50.17 STORM can not only be used for
00:22:50.18 00:22:51.24 imaging fixed samples,
00:22:51.25 00:22:54.28 but can also be used for live cell imaging.
00:22:54.29 00:22:57.12 Here I show you two examples,
00:22:57.13 00:23:01.13 and the first is actually a three dimensional, two color
00:23:01.14 00:23:05.07 STORM image of Clathrin coated pits,
00:23:05.08 00:23:07.24 and its cargo transferrin in living cells.
00:23:07.25 00:23:11.25 Here you can see the transferrin cargo inside the pit
00:23:11.26 00:23:13.23 and here you can see the transferrin
00:23:13.24 00:23:15.24 actually moving towards the pit.
00:23:15.25 00:23:17.05 And in the second example,
00:23:17.06 00:23:21.02 is a STORM movie showing you
00:23:21.03 00:23:23.19 fission and fusion of mitochondria.
00:23:23.20 00:23:26.10 What's quite remarkable about this movie
00:23:26.11 00:23:29.13 is that it is actually taken with the sample
00:23:29.14 00:23:31.27 just simply labeled with MitoTracker,
00:23:31.28 00:23:34.27 which is one of the very easy to use,
00:23:34.28 00:23:37.21 and commonly used probes that people use to
00:23:37.22 00:23:41.19 label mitochondria in living cells.
00:23:41.20 00:23:43.20 And then here you can clearly see
00:23:43.21 00:23:45.26 the fission and fusion intermediate,
00:23:45.27 00:23:47.12 the tubular type of structure.
00:23:47.13 00:23:50.12 So typically we can accomplish
00:23:50.13 00:23:55.15 a spatial resolution about 20 to 40 nm,
00:23:55.16 00:23:58.25 and a temporal resolution on the order of 1 sec,
00:23:58.26 00:24:01.09 and these images are taken typically
00:24:01.10 00:24:03.20 with 1 kHz type of frame rate.
00:24:03.21 00:24:09.09 So 1000 images to give you one movie.
00:24:09.10 00:24:14.15 STORM can not only be applied to image cultured cells,
00:24:14.16 00:24:18.29 it can also be applied to image tissue samples.
00:24:19.00 00:24:21.25 For example, what I'm showing here is a
00:24:21.26 00:24:24.20 collaborative work between Catherine Dulac's lab
00:24:24.21 00:24:27.18 at Harvard and my lab, where we
00:24:27.19 00:24:30.29 imaged synapses in brain tissue slices.
00:24:31.00 00:24:33.29 And what you see here is a conventional
00:24:34.00 00:24:38.12 wide field image of synapses labeled with
00:24:38.13 00:24:41.03 Bassoon, a presynaptic scaffolding protein,
00:24:41.04 00:24:44.18 and Homer1, a postsynaptic scaffolding protein.
00:24:44.19 00:24:48.21 And here's the STORM image of the same field of view.
00:24:48.22 00:24:51.29 Since this is a three dimensional image,
00:24:52.00 00:24:56.03 we can sort of take one of these synapses and
00:24:56.04 00:24:58.17 show you a three dimensional perspective
00:24:58.18 00:24:59.18 showing the movie
00:24:59.19 00:25:02.04 where you can clearly see the pancake shape
00:25:02.05 00:25:06.11 like structure for the presynaptic active bassoon
00:25:06.12 00:25:07.24 and the post synaptic density
00:25:07.25 00:25:12.01 and also the clear separation of these two structures
00:25:12.02 00:25:14.17 by the synaptic cleft.
00:25:14.18 00:25:19.11 You will hear more about this single molecule based
00:25:19.12 00:25:23.05 superresolution fluorescent microscopy approach
00:25:23.06 00:25:26.24 from Jeniffer Lippincott-Schwartz, and Bo Huang
00:25:26.25 00:25:30.19 who both will speak in this iMicroscopy lecture series.
00:25:30.20 00:25:34.27 So I'll also not dwell in the technicality
00:25:34.28 00:25:37.05 of this approach any further.
00:25:37.06 00:25:40.28 But I will end this lecture with one recent
00:25:40.29 00:25:43.24 biological application where we
00:25:43.25 00:25:46.05 applied STORM to study
00:25:46.06 00:25:50.29 actin cytoskeleton structure in neurons.
00:25:51.00 00:25:53.03 So as we all know
00:25:53.04 00:25:57.10 actin play a variety of very important roles in cells,
00:25:57.11 00:26:00.00 and also in particular in neurons.
00:26:00.01 00:26:04.17 It is important for the polarization of neurons.
00:26:04.18 00:26:09.09 It is important for the stabilization of axons.
00:26:09.10 00:26:11.20 It is important for the stabilization
00:26:11.21 00:26:13.26 and plasticity of synapses.
00:26:13.27 00:26:17.28 And it is also important for intracellular transport
00:26:17.29 00:26:20.11 and other functions.
00:26:20.12 00:26:23.08 So because of these important functional role
00:26:23.09 00:26:26.17 it will be really nice if we know how actin filaments
00:26:26.18 00:26:30.15 are organized, how actin filaments are organized in
00:26:30.16 00:26:33.26 neurons, but very little is known about that.
00:26:33.27 00:26:36.07 In particular, very little is known
00:26:36.08 00:26:38.18 about what actin filament
00:26:38.19 00:26:41.17 is organized like in axonal shaft,
00:26:41.18 00:26:43.29 and in dendritic shaft.
00:26:44.00 00:26:48.06 And that again is because what I said earlier
00:26:48.07 00:26:51.14 that actin filaments are so skinny,
00:26:51.15 00:26:53.05 and they are packed in high density,
00:26:53.06 00:26:54.27 they are packed in high density together
00:26:54.28 00:26:56.20 with other filamentous structures,
00:26:56.21 00:26:58.07 in particular in axons,
00:26:58.08 00:26:59.19 they are packed together with
00:26:59.20 00:27:01.14 neurofilaments and microtubules.
00:27:01.15 00:27:04.22 So because of that, to resolve actin structure
00:27:04.23 00:27:10.16 actually requires both ultra high spatial resolution
00:27:10.17 00:27:13.03 and extremely high molecular specificity.
00:27:13.04 00:27:16.14 And therefore superresolution fluorescent microscopy
00:27:16.15 00:27:19.17 provides ideal approach to solve this problem.
00:27:19.18 00:27:23.05 So here what you see is a three dimensional
00:27:23.06 00:27:28.29 STORM image of actin in the dendritic compartment.
00:27:29.00 00:27:31.00 We know it's the dendritic compartment
00:27:31.01 00:27:35.09 because the red label here, actually the MAP2,
00:27:35.10 00:27:37.06 is dendrite specific.
00:27:37.07 00:27:40.25 And these Phalloidin labeled actin filament,
00:27:40.26 00:27:43.04 you can see that they run along
00:27:43.05 00:27:46.23 the longer axis of the dendritic shaft.
00:27:46.24 00:27:49.14 This is a type of organization
00:27:49.15 00:27:51.00 that is not so hard to imagine.
00:27:51.01 00:27:54.04 And what is very surprising to us
00:27:54.05 00:27:56.04 is when we look at
00:27:56.05 00:27:59.16 what actin look like in axonal shafts,
00:27:59.17 00:28:02.16 and we see this drastically different picture.
00:28:02.17 00:28:06.27 Instead of running along the long axis of axons,
00:28:06.28 00:28:08.27 which we do see a few filaments like that,
00:28:08.28 00:28:12.28 but the most prominent feature or aspect is
00:28:12.29 00:28:17.09 this stripes, these stripes that are perpendicular
00:28:17.10 00:28:19.20 to the long axis of the axons.
00:28:19.21 00:28:23.06 And they are arranged in this very regular pattern.
00:28:23.07 00:28:25.25 And it is very hard to see by other means,
00:28:25.26 00:28:26.16 for example,
00:28:26.17 00:28:28.25 this is a conventional fluorescence image
00:28:28.26 00:28:30.07 of the same field of view.
00:28:30.08 00:28:32.14 Now because this is a 3D image,
00:28:32.15 00:28:35.24 we can also look at what these stripes are by
00:28:35.25 00:28:37.16 turning the thing aside,
00:28:37.17 00:28:38.24 and you can see they are
00:28:38.25 00:28:41.19 actually ring-like structures
00:28:41.20 00:28:44.27 that wrap around the circumference of the axons.
00:28:44.28 00:28:48.00 As I said, they are highly regular.
00:28:48.01 00:28:51.11 You can take any segment of axons,
00:28:51.12 00:28:54.01 and then project the localization,
00:28:54.02 00:28:55.04 the actin localization
00:28:55.05 00:28:56.09 into one dimension,
00:28:56.10 00:28:59.06 and then you can subject this one dimensional
00:28:59.07 00:29:01.26 localization distribution to Fourier analysis,
00:29:01.27 00:29:05.12 and then see this clearly peaky structures.
00:29:05.13 00:29:07.02 You can see the primary peak,
00:29:07.03 00:29:09.19 and the secondary and tertiary hormonic and so on.
00:29:09.20 00:29:12.02 And then for this particular segment,
00:29:12.03 00:29:14.29 these Fourier analysis, the peak
00:29:15.00 00:29:19.18 corresponds to about 190 nm periodicity.
00:29:19.19 00:29:22.07 And then if we look at many filaments,
00:29:22.08 00:29:24.06 this is the distribution,
00:29:24.07 00:29:31.01 many axons, this is the distribution of periodicity
00:29:31.02 00:29:36.09 which is about 182 nm +- 16 nm.
00:29:36.10 00:29:42.00 So how do the actin filament manage to organize
00:29:42.01 00:29:44.03 into such a periodic structure,
00:29:44.04 00:29:45.05 in other words,
00:29:45.06 00:29:50.05 what are the factors that are spacing the actin rings.
00:29:50.06 00:29:54.00 So the spacing periodicity of 182-190 nm
00:29:54.01 00:29:56.07 actually offered us some clue.
00:29:56.08 00:30:00.12 Because that reminded us of this molecular complex
00:30:00.13 00:30:02.05 called spectrin tetramers.
00:30:02.06 00:30:04.24 And spectrin tetramer,
00:30:04.25 00:30:10.11 the only place where the organization of spectrin
00:30:10.12 00:30:12.16 is known is actually in red blood cells.
00:30:12.17 00:30:16.20 And there, spectrin tetramers form these triangular
00:30:16.21 00:30:20.25 shaped network, or polygonal shaped network,
00:30:20.26 00:30:23.15 where the spectrin tetramer forms the edge,
00:30:23.16 00:30:27.01 and at the nodes are actually short actin filaments.
00:30:27.02 00:30:28.28 So in other words, you can see that spectrin
00:30:28.29 00:30:32.16 can connect neighboring short actin filaments.
00:30:32.17 00:30:35.21 And spectrin is ubiquitously expressed in
00:30:35.22 00:30:37.29 many metazoan cells.
00:30:38.00 00:30:41.13 And therefore, it is also in the brain.
00:30:41.14 00:30:45.08 And actually brain spectrin purified
00:30:45.09 00:30:48.19 has a length of about 195 nm.
00:30:48.20 00:30:52.01 So you put these together, then spectrin becomes
00:30:52.02 00:30:54.14 a very good candidate to be the one that's spacing
00:30:54.15 00:30:57.02 actin filaments, because it can connect actin
00:30:57.03 00:30:58.24 and it's about 190 nm long.
00:30:58.25 00:31:01.21 So that led us to this hypothesis,
00:31:01.22 00:31:03.21 that the ring we see are
00:31:03.22 00:31:05.24 actually short actin filaments
00:31:05.25 00:31:08.00 and we think they are short because
00:31:08.01 00:31:10.29 in analogy to red blood cell.
00:31:11.00 00:31:13.07 And we think they are capped by adducin,
00:31:13.08 00:31:17.26 again in a analogy to the red blood cell structure.
00:31:17.27 00:31:19.26 And these structure of actin filament line up into
00:31:19.27 00:31:22.12 this kind of ring structure wrap around
00:31:22.13 00:31:24.06 the circumference of the axon.
00:31:24.07 00:31:27.07 And neighboring actin rings are actually
00:31:27.08 00:31:30.23 spaced and connected by spectrin tetramers.
00:31:30.24 00:31:33.03 Now to test that,
00:31:33.04 00:31:37.15 then we can image spectrin and adducin
00:31:37.16 00:31:39.01 together with actin.
00:31:39.02 00:31:41.17 For example, if we image the middle portion of
00:31:41.18 00:31:43.16 the spectrin tetramer, for example,
00:31:43.17 00:31:46.21 we can immunostain the C-terminal domain
00:31:46.22 00:31:49.00 of beta two spectrin, which is in the middle part
00:31:49.01 00:31:50.04 of the spectrin tetramer,
00:31:50.05 00:31:53.06 then we should see that stain to alternate with
00:31:53.07 00:31:54.05 the actin stain,
00:31:54.06 00:31:56.28 forming these alternating stripe pattern,
00:31:56.29 00:31:59.21 which is indeed what we see in the two color
00:31:59.22 00:32:01.27 STORM image of actin and the spectrin
00:32:01.28 00:32:04.07 C terminal domains.
00:32:04.08 00:32:07.19 And adducin we expected to line up with actin,
00:32:07.20 00:32:09.14 which is indeed also what we see.
00:32:09.15 00:32:11.00 And as a corollary,
00:32:11.01 00:32:13.19 you can see that the spectrin C terminal domain
00:32:13.20 00:32:17.17 and adducin form those alternating patterns.
00:32:17.18 00:32:21.11 This is a quite prevalent structure,
00:32:21.12 00:32:24.24 here is an entire field of view we have taken
00:32:24.25 00:32:28.13 for hippocampal cultured neuron,
00:32:28.14 00:32:31.07 and then you can see that except for this
00:32:31.08 00:32:33.07 big, fat dendrite in the middle,
00:32:33.08 00:32:36.29 all the other axonal structures show this
00:32:37.00 00:32:40.17 periodic, ladder-like pattern.
00:32:40.18 00:32:43.21 And this sample is stained for Beta-II spectrin.
00:32:43.22 00:32:47.04 And we can not only see this in cultured neurons,
00:32:47.05 00:32:49.27 we can also see it in brain tissue slices.
00:32:49.28 00:32:53.05 And here is the spectrin stain in brain
00:32:53.06 00:32:55.19 tissue slices where you can clearly see
00:32:55.20 00:32:57.06 the periodic-like pattern.
00:32:57.07 00:32:59.11 So put it all together,
00:32:59.12 00:33:04.06 we found that in axons, there is this novel
00:33:04.07 00:33:06.26 submembrane cytoskeletal structure that
00:33:06.27 00:33:08.12 was previously unknown.
00:33:08.13 00:33:11.24 And this structure is formed by
00:33:11.25 00:33:14.08 periodically arranged actin rings.
00:33:14.09 00:33:15.13 And these actin rings are
00:33:15.14 00:33:18.24 short filaments of actin alined around the
00:33:18.25 00:33:21.12 circumferential direction of the axons,
00:33:21.13 00:33:23.01 and these neighboring actin rings are
00:33:23.02 00:33:26.07 connected by spectrin tetramers.
00:33:26.08 00:33:31.09 So we have also results, and reasons to
00:33:31.10 00:33:35.04 believe that this periodic remarkably
00:33:35.05 00:33:37.03 long range ordered structure
00:33:37.04 00:33:39.26 plays some important roles for neurons.
00:33:39.27 00:33:43.13 For example, because spectrin is flexible,
00:33:43.14 00:33:46.01 so this cytoskeletal structure is
00:33:46.02 00:33:47.27 both robust and flexible,
00:33:47.28 00:33:50.22 and it can provide the kind of elastic and
00:33:50.23 00:33:53.11 durable mechanical support that
00:33:53.12 00:33:58.22 a thin and long axon filament needs.
00:33:58.23 00:34:03.00 And also, because this is a submembrane
00:34:03.01 00:34:04.27 lattice structure, it actually can
00:34:04.28 00:34:07.04 anchor membrane proteins
00:34:07.05 00:34:09.05 also in a periodic pattern.
00:34:09.06 00:34:11.01 And for example, we know
00:34:11.02 00:34:12.11 some of the very important
00:34:12.12 00:34:13.19 functional membrane proteins
00:34:13.20 00:34:16.21 in axons, like sodium channels,
00:34:16.22 00:34:18.15 we also found that it also forms a periodic
00:34:18.16 00:34:21.22 pattern in accordance with the underlying
00:34:21.23 00:34:23.08 cytoskeletal structure.
00:34:23.09 00:34:28.11 So this new structure could potentially
00:34:28.12 00:34:33.28 be important, novel structure for neurons.
00:34:33.29 00:34:37.12 But again, because of the time limit of this
00:34:37.13 00:34:40.04 short intro lecture, I won't elaborate
00:34:40.05 00:34:43.17 further on the structure, the development
00:34:43.18 00:34:47.27 function of this cytoskeleton structure.
00:34:47.28 00:34:54.00 But hopefully, through these brief introductions
00:34:54.01 00:34:56.28 of the variety of superresolution fluorescence
00:34:56.29 00:34:59.15 microscopy methods, and through this
00:34:59.16 00:35:02.08 one interesting recent biological application,
00:35:02.09 00:35:06.02 I hope I have conveyed you this message
00:35:06.03 00:35:09.23 that these new fluorescent imaging methods
00:35:09.24 00:35:12.10 are indeed powerful, and they can indeed
00:35:12.11 00:35:15.10 lead to new biological discoveries.
00:35:15.11 00:35:19.03 And this a very rapidly developing field
00:35:19.04 00:35:22.19 with many new things, new methods,
00:35:22.20 00:35:26.10 and discoveries that keep coming out.
00:35:26.11 00:35:28.26 For those interested audience,
00:35:28.27 00:35:32.10 you can also find exciting lectures
00:35:32.11 00:35:37.09 given by Stefan Hell, Jeniffer Lippincott-Schwartz,
00:35:37.10 00:35:38.20 Dave Agard and Bo Huang
00:35:38.21 00:35:41.25 about superresolution fluorescence microscopy
00:35:41.26 00:35:45.10 and in this same microscopy lecture series.
00:35:45.11 00:35:47.03 So with that, let me thank
00:35:47.04 00:35:49.00 the people in my lab
00:35:49.01 00:35:51.20 who both former group members and
00:35:51.21 00:35:53.15 current group members who have contributed
00:35:53.16 00:35:55.16 to this work, in particular,
00:35:55.17 00:35:59.19 to the studies I discussed in this lecture.
00:35:59.20 00:36:03.00 And the early development of STORM
00:36:03.01 00:36:07.20 is largely due to three former group members
00:36:07.21 00:36:10.27 Bo Huang, Mark Bates, and Mike Rust.
00:36:10.28 00:36:14.03 And some of the more recent development of
00:36:14.04 00:36:15.24 the STORM I have talked here,
00:36:15.25 00:36:19.01 for example, the live cell STORM imaging
00:36:19.02 00:36:21.15 is due to Sang-Hee Shim,
00:36:21.16 00:36:24.16 and Sara Jones, and the improvement
00:36:24.17 00:36:28.23 of resolution by the dual objective geometry
00:36:28.24 00:36:34.28 is due to Ke Xu and Hazen Babcock.
00:36:34.29 00:36:38.18 And the biological application of studying
00:36:38.19 00:36:42.04 actin and spectrin dependent cytoskeletal structure
00:36:42.05 00:36:48.07 in neurons is done by Ke Xu and Guisheng Zhong.
00:36:48.08 00:36:51.06 And I would also like to thank
00:36:51.07 00:36:53.26 Howard Huge Medical Institute and
00:36:53.27 00:36:55.11 the National Institutes of Health
00:36:55.12 00:36:58.06 for their generous support.
00:36:58.07 00:37:02.19 And also very importantly, thank you
00:37:02.20 00:37:05.26 the captivating audience for being interested in
00:37:05.27 00:37:07.27 advanced microscopy approaches.
00:37:07.28 00:37:32.27 Thank you!

This Talk

Speaker: Xiaowei Zhuang

Audience:

Student
Educators of Adv. Undergrad / Grad
Researcher
Educators

Recorded: July 2013

Talk Overview

This lecture surveys a variety of recent methods that achieve higher resolution than is possible with conventional microscopy with diffraction-limited optics. These include different types of patterned illumination (e.g. STED and SIM microscopy) or techniques that build up an image by stochastically switching on single fluorescent molecules and localizing each molecule with high spatial precision (STORM, PALM, FPALM).

Questions

What is the precision at which one could localize the position of single fluorescent dye molecule if 10,000 photons are collected and the width of the diffraction limited spots is 250 nm?
Match the following descriptions to the indicated super-resolution technique
1. STED/RESOLFT
2. SIM (structured illumination microscopy)
3. PALM/F-PALM
4. STORM
  i. Uses a sinusoidal illumination pattern to enhance high resolution information from the specimen by 2-3-fold
  ii. High precision localization of well separated single, photoactivatable GFP molecules, which through iterations of photoactivation can build up an image.
  iii. Uses a second donut shaped illumination beam to deplete fluorescence and reduce the point spread function.
  iv. High precision localization of photoswitchable organic fluorophores which can be activated/deactived by light.
STORM/PALM or SIM: which technique typically has faster temporal resolution?
STORM/PALM or SIM: which technique typically has better spatial resolution?

Answers

View Answers

Speaker Bio

Xiaowei Zhuang

Xiaowei Zhuang is Professor of Chemistry and Chemical Biology, and Professor of Physics at Harvard University and a Howard Hughes Medical Institute Investigator. She received her B.S. in physics from University of Science and Technology of China and her Ph.D. in physics from the University of California, Berkeley. As a post-doctoral fellow with Steven Chu… Continue Reading